Software for a continuous machining production system with machining centers was debugged by the use of L36 orthogonal arrays, based on the probability matching method. A combination of five orthogonal arrays was used, roughly classified into operating conditions, detailed settings, and uncertain factors. With conventional verificatien methods, it took several weeks to find ordinaly bugs; with the new method, verification was completed in two days. Since then,no bugs have been found in the delivered software for several years except for one, which was purely external. Introduction of the new method has improved software reliability and has been economically effective as well.

In inductive heating, eddy currents are generated when an alternating magnetic flux created by application of alternating current to an exciting coil passes through metal, and the eddy current loss becomes Joule's heat, Compared with radiative heating by a halogen lamp, which is conventionally employed in the fusing sections of copiers, inductive heating should be more efficient because it generates heat directly in the heated body. In the present study, several conceptual systems that emerged from the exploratory phase of a project to use inductive heating for fusing were narrowed down to a single system by using CAE to perform a parameter design of each system and comparing the systems in terms of their S/N ratios. In addition to the comparison ef optimum conditions obtained through CAE, prototypes with near-optimum parameters were built and their performance was evaluated. The results showed excellent agreement with the results of the CAE parameter design. This method of system selection by in-depth exploration through parameter design is proposed as a suitable method for system selection in studies aimed at the development of new technology.

In recent years, evaluations of electric power consumption as a fundamental property of cutting and other machining systems, based on the energy conversion concept, are beginning to produce many promising results. This study presents a case of successful application to the end milling of a high-hardness material. The test pieces were flat plates, with differences in stiffness due to differences in plate thickness as a noise factor. The process was evaluated in two ways: from the amount of material removed versus number of passes, and from electric power consumed versus number of passes. Except for the signal-to-noise ratio calculated from amount of material removed versus number of passes, the estimated gain closely matched the actual gain, demonstrating reproducibility of the results. Analysis of the relationship between the number of passes M' and the amount of material removed M produced the correct amount of material removed per pass, and identified the conditions for cutting with a high removal rate, which was the goal of the experiment.

Through the process of using the MT system to determine changes in states of human fatigue as part of a study of foot-bathing effect, it was found that human fatigue can be quantified. Based on this result, the MT method was used to determine fatigue distance changes from an L108 orthogonal array and find trends in changes in state of fatigue, taking into account a total of 110 factors,including seven factors such as day, time, and age, twenty-four physical factors,seventeen lifestyle factors, and sixty-two factors of the type found on Kampo diagnostic questionnaires. Relationships between the factors and everyday fatigue changes were also studied by diagnosing the factors separately. Many of the fatigue changes related with individual factors can be predicted from experience, but fatigue changes that are not necessarily easy to predict were measured. Fatigue changes due to individual factors, however, were minimal. In terms of distance, the effect of a single factor is only 0.05, equivalent to the effect of climbing or descending an eight-step flight of stairs. It is numerical demonstrated that fatigue arises from complex interactions of many such factors.